UCLA Professors Say eHarmony Is Unscientific and Its Customers Are 'Duped.' Here's Why.
The biggest problem with Warren's algorithm is that it seems to be based on conclusions drawn from already married couples: He says that similar people are more likely to form happy, long-term relationships. But Karney points out that successful couples tend to perceive themselves as similar, regardless of whether they would have done so as unacquainted strangers. "If I like you, I'll find a way to be similar," he says. So perceived similarities are a consequence, not a cause, of strong relationships.
Perhaps to head off scrutiny once eHarmony grew to millions of members, Warren established a research facility in 2007. The senior director of research & development at eHarmony Labs, Gian Gonzaga, is also an adjunct professor at UCLA, where Bradbury served as his post-doc supervisor.
For eHarmony, affiliation with Gonzaga and Bradbury appears to be PR gold that creates the impression that the site's matching methods are based on hard science. The eHarmony Labs website claims Bradbury and other advisers "work collaboratively with us to develop eHarmony products... emphasiz[ing] eHarmony Lab's commitment to ... providing research-driven products." But according to Bradbury, his advice has no influence on eHarmony's actual product: the matching algorithm developed by Warren in the late 1990s.
The question remains whether the legitimate research Gonzaga has been churning out will be used for anything more than publicity.
"Tom and I could easily design a study," Karney says, "and eHarmony has the resources ... that would prove, 'Look, when we put [users] through our algorithm, they do better than when you put them through another algorithm or a control condition.'"
"This is one of the studies that I told them to do," Bradbury says. "Five years ago."
"That's a nightm--" Karney starts to say, trying to shape his look of disbelief into a smile. "See, that's why I'm delighted that he's on the board. I'm sure that he's only told them smart things. You know, he's a smart guy," he stage-whispers.
With Bradbury on eHarmony's payroll but incapable of directly influencing the "science" behind its product, Karney is left to fight openly for scientific integrity and take down eHarmony.
Except that's not what he wants.
"Has your group gone to a regulatory agency and said, 'We are outraged at the misuse of our science?'" Bradbury asks.
"First of all, outrage is a little strong. I didn't say I was outraged," Karney says.
The conclusion section of the new study advocates "closer collaboration between scholars and service providers" and seems to be demanding not increased regulation but a piece of the action. The section lauds cooperation between academics and matchmakers as "an unprecedented opportunity for researchers to test their theories and develop new ones with large samples of participants," which translates into an offer of help with the heavy science lifting in exchange for access to the 21st century's most valuable currency: information, specifically the gold mine of user data collected by dating sites.
"Why spend all this time and energy if you don't want to change things?" Bradbury asks, prodding his critical colleague.
"I think it matters, I agree," Karney says with a sheepish smile. "We haven't gone to a regulatory agency, only because I've been a little busy this week."